Weather Chatbot

A weather chatbot built using Llamaindex and OpenWeatherMap APIs. This chatbot is capable of answering questions about the weather, comparing forecasts, and finding the closest cities to specified coordinates. It runs locally on your machine using Streamlit.

Features

• City Weather Queries: Get the current weather for any city by name.
• Coordinate-Based Weather: Retrieve weather information based on geographic coordinates (latitude, longitude).
• Weather Forecasting: Provides a detailed 5-day forecast with 3-hour intervals using OpenWeatherMap API.
• Weather Comparison: Compare the weather between multiple cities.
• Nearest Cities Search: Find the closest cities to a given set of geographic coordinates.
• Geocoding/Reverse Geocoding: Convert between city names and geographic coordinates.
• Flexible Data Access: Leverage a CSV of cities for efficient querying, enabling fast, localized intermediate processing and responses.

Tools Included

1. Geography: This tool uses the PandasQueryEngine to interact with a CSV file of cities supported by OpenWeather. The LLM infers data frame operations to retrieve the result.
   ⚠️ Warning: This tool allows the LLM access to the eval function, which can lead to arbitrary code execution on the machine running this tool. Use with caution.

2. Closest_cities: This tool returns the nearest cities along with their latitude and longitude based on the provided coordinates.

3. OpenWeatherMap API Tools:

   • current_weather: Retrieves the current weather for a set of coordinates.
   • Day5_hour3_forecast: Provides a 5-day weather forecast for a set of coordinates with 3-hour intervals.
   • geocoding and reverse_geocoding: Converts between geographic coordinates and city names.

Setup

1. Clone the repository.

2. Create a file named creds and add the following keys:

   • open_ai_key: Your OpenAI API key.
   • OpenWeatherMap API key: Your OpenWeatherMap API key.
   • csv_filepath: The path to the merged_locations.csv file.

   Example creds file:

   open_ai_key=YOUR_OPENAI_API_KEY
   OpenWeatherMap API key=YOUR_OPENWEATHER_API_KEY
   csv_filepath=./data/merged_locations.csv
   

3. Run the Streamlit application.

   streamlit run main.py

Static Data Sources

• City list from OpenWeatherMap: city_list
• Country codes: country_codes

These data sources were merged and cleaned to create the merged_locations.csv file, which the chatbot uses to answer location-based weather questions.

Example Screenshots

Here are some screenshots of the chatbot in action:

License

MIT License

Feel free to contribute to the project and enhance the chatbot's capabilities!